Here are a few sentences which you might have a view on:
1. ‘The vixens are all and only the female foxes’ (analytic)
2. ‘Superman is Superman’ (analytic)
3. ‘Clark Kent is Superman’ (not analytic)
4. ‘Water is H2O’ (not analytic)
5. ‘Gorse is furze’ (?)
6. ‘The women are all and only the female adult humans’ (analytic)
7. ‘The ladies are all and only the female adult humans’ (analytic)
8. ‘Women are ladies’ (?)
(1) is meant to be a paradigm case, and (6) and (7) aren’t relevantly different from it. The difference between (2) and (3) is meant to bring out the difference between informative and uninformative identities which Frege and I have spent time fretting about. (4) is meant to be like (3) but with a natural kind instead of a superhero. (5) is meant to demand the same verdict as (4) but be more similar to (8). (If you’re not persuaded that (5) isn’t analytic but you like Putnam, then consider what he said about beeches and elms.) But if (6) and (7) entail (8), (8) should be analytic too.
What’s happened? Obviously I think that what’s happened is the distinction’s bunk. I’m not sure what the proper response for my opponents is. I think one thing they’ll say is that the concepts woman and lady are both parasitic on the same concept human. But if that’s true, then I don’t see that there are two distinct concepts in the gorse/furze case either: just one concept that you could acquire twice without realising it. So we’ll have to say that ‘gorse is furze’ is analytic, but if we say that, we can’t use analyticity to explain why it’s informative and ‘gorse is gorse’ isn’t. While you’re here, try these:
9. ‘The women are all and only the female adult homo sapiens’ (?)
10. ‘Female adult humans are female adult homo sapiens’ (?)
11. ‘Humans are in the genus homo’ (not analytic)
12. ‘Homo sapiens are in the genus homo’ (analytic)
(10) follows from (6) and (9). I’d have thought (9) looked pretty analytic, but (10) shouldn’t be analytic if my verdicts on (11) and (12) are right. I’m not sure how persuasive any of this should be. I expect the thing to do is accuse me of equivocation here and there, but I don’t know how much mileage there would be in that, and in any case if 'H2O' and 'homo sapiens' are ambiguous then that's interesting in itself. If I've done something straightforwardly foolish I've also got a thought-experiment about an invasion of bachelors, but I won't get into that here.
Essentially I’m trying to drive a wedge between saying that true identity statements with rigid designators on both sides are ever synthetic, and saying paradigms like ‘vixens are female foxes’ are analytic. If you deny the paradigms you’ve conceded a lot, and if you deny synthetic identities you can’t use analyticity to explain why some identities are informative and some aren’t. If you’re not careful, analyticity won’t be able to explain why anything is a priori. A concept of analyticity as weak as that might be one I could get on board with.